English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Federated Learning (FL) addresses critical issues in machine learning relatedto data privacy and security, yet suffering from data insufficiency andimbalance under certain circumstances. The emergence of foundation models (FMs)offers potential solutions to the limitations of existing FL frameworks, e.g.,by generating synthetic data for model initialization. However, due to theinherent safety concerns of FMs, integrating FMs into FL could introduce newrisks, which remains largely unexplored. To address this gap, we conduct thefirst investigation on the vulnerability of FM integrated FL (FM-FL) underadversarial threats. Based on a unified framework of FM-FL, we introduce anovel attack strategy that exploits safety issues of FM to compromise FL clientmodels. Through extensive experiments with well-known models and benchmarkdatasets in both image and text domains, we reveal the high susceptibility ofthe FM-FL to this new threat under various FL configurations. Furthermore, wefind that existing FL defense strategies offer limited protection against thisnovel attack approach. This research highlights the critical need for enhancedsecurity measures in FL in the era of FMs.

Vulnerabilities of Foundation Model Integrated Federated Learning Under Adversarial Threats